Mitochondrial DNA (mtDNA) is highly susceptible to mutations due to the low level of DNA repair and the presence of a high level of reactive oxygen species in the organelle. Although mtDNA mutations have been implicated in degenerating diseases, aging, and cancer, very little is known about the role of T cells in immunosurveillance for mtDNA aberrations. Here, we describe T-cell recognition of a peptide translated from an alternative open reading frame of the mitochondrial cytochrome b (cyt b) gene in melanoma cells established from a patient. To understand how the cyt b gene is transcribed and translated in tumor cells, we found that cyt b–specific CD4+ T cells only recognized protein fractions derived from cytoplasm and not from mitochondria. However, T-cell recognition of tumor cells could be inhibited by treatment of tumor cells with rhodamine 6G inhibitor, which depletes mitochondria. These findings suggest that cyt b mRNA is leaked out of the mitochondria and then translated in the cytoplasm for presentation to CD4+ T cells. The cyt b cDNAs from this patient contain highly heteroplasmic transition mutations compared with control cell lines, suggesting a compromise of mitochondrial integrity that may have contributed to melanoma induction or progression. These findings provide the first example of a mitochondrial immune target for CD4+ T cells and therefore have implications for the immunosurveillance of mitochondrial aberrations in cancer patients. (Cancer Res 2006; 66(11): 5919-26)

CD4+ helper T cells play important roles in orchestrating the host immune response against cancer and infectious diseases (1). Recent studies show that such cells are required for priming as well as secondary expansion of CD8+ memory T cells (24). They can also mediate tumor regression in the absence of CD8+ T cells, including MHC class II–negative or MHC class II–positive tumors (5, 6). Thus, the identification of MHC class II–restricted tumor antigens capable of stimulating CD4+ helper T cell responses has been critical for the development of effective cancer vaccines.

Over the past few years, a limited number of MHC class II–restricted tumor antigens have been identified in melanoma and other types of cancers (1) and have been classified into four major groups based on their expression patterns. The tissue-specific differentiation antigens, such as tyrosinase and glycoprotein 100, are expressed in melanoma as well as normal melanocytes. Tumor-specific shared antigens, such as LAGE-1 and NY-ESO-1, are expressed in many tumors but not in normal tissues, except testis. Antigens that are overexpressed in tumors include telomerase and MUC1. Tumor-specific unique or mutated antigens, such as CDC27 and LDLR-fucosyltransferase, result from point mutations and genetic alterations. These MHC class II–restricted and MHC class I–restricted tumor-associated antigens are derived from proteins localized in different subcellular compartments, including nucleus, cytosol, melanosomes, or plasma membrane (710). Although mitochondrial antigens, such as PDC-E2 (11, 12), aconitase (13), NubO (14), and ND1 (15), originated from nucleus and mitochondrion have been reported to be recognized by CD4+ and CD8+ T cells in various diseases, very little is known about mitochondrial antigens recognized by tumor-specific T cells. Because aberrant mitochondrial proteins are frequently observed in a variety of cancers, including melanoma (16), we postulate that some of these antigens may trigger an immune response.

Here, we describe the identification and characterization of a mitochondrially derived antigen, cytochrome b (cyt b), as a new MHC class II–restricted tumor antigen recognized by tumor-reactive CD4+ T cells. The T-cell epitope was derived from an alternative open reading frame (ORF) of the cyt b gene. These novel findings suggest a new role of CD4+ T cells in immunosurveillance for mitochondrial aberrations in cancer patients.

Reagents, cell lines, and cultures. Antibodies used in the histocompatibility leukocyte antigen (HLA)-blocking experiments were described previously (17). Anti-CD4 and anti-CD8 antibodies used for the fluorescence-activated cell sorting (FACS) analysis of CD4+ T cells were purchased from BD PharMingen (San Diego, CA). MHC class II pathway inhibitor chloroquine and mitochondrion inhibitor rhodamine 6G were purchased from Sigma-Aldrich (St. Louis, MO). Melanoma cell lines, EBV-transformed B-cell lines, and HEK293IMDR1 cells were established in our laboratory and maintained in RPMI 1640 (Biofluids, Gaithersburg, MD) supplemented with 10% FCS (Biofluids). CD4+ 1362 tumor-infiltrating lymphocytes (TIL) were cultured from a s.c. metastases resected from a patient (no. 1,362). T-cell lines were grown in RPMI 1640 containing 10% human AB serum (Valley Biochemicals, Winchester, VA) and 300 IU/mL interleukin 2 (IL-2; Chiron Corp., Emeryville, CA). T-cell sublines were generated by limiting dilution methods from a well plated at three cells per well as described previously (18).

cDNA library construction and screening. Total RNA was isolated from 1362mel cells using the RNeasy system (Qiagen, Valencia, CA). Polyadenylated RNA [poly(A)+ RNA] was purified from total RNA by the PolyATract system (Promega, Madison, WI). Approximately 5 μg poly(A)+ RNA was converted to cDNA using a cDNA construction kit (Life Technologies, Rockville, MD) with an oligo(dT) primer bearing a NotI site. Construction and screening of the cDNA library of 1362mel tumor cell RNA were similar to those described previously (10, 19). Plasmid DNAs containing cyt b fragment were sequenced and assembled with SEQUENCHER software (20).

Subcellular organelle fractionation, tumor cell lysate preparation, and peptide pulsing. EBV-transformed B cells were used as the antigen-presenting cells (APC) for the cell lysate and peptide-pulsing experiments. Subcellular fractionation of organelles from 586mel and 1362mel cells was done using sucrose gradient centrifugation approaches as described previously (21). Organelle lysate from fractionated organelles was obtained by repeated (thrice) freezing and thawing in the presence of 0.1% Triton X-100/PBS. Total cell lysate was prepared by repeated (thrice) freezing and thawing of tumor cells in PBS at ∼2 × 107/mL followed by brief centrifugation at a low speed to pellet the unbroken cells and large cell debris. Protein concentration from individual fraction was determined by the bicinchoninic acid assay (Pierce Endogen, Rockford, IL), and an equal amount of protein was pulsed onto 1362 EBV-transformed B cells and incubated overnight. Pulsed EBV-transformed B cells were washed thrice before coculture with 5 × 104 CD4+ T cells. cyt b synthetic peptides used for peptide titration experiments were pulsed onto EBV-transformed B cells at a various concentrations for 4 hours in RPMI 1640, washed twice with RPMI, and then cocultured with CD4+ T cells overnight for cytokine release assays.

Construction of cyt b–specific short hairpin RNA lentiviral vectors. Plasmid PLL3.7 was used to insert cyt b–specific short hairpin RNA (shRNA) oligos. The selection of coding sequences for shRNAs was empirically determined but started with an AG or AAG and analyzed by BLAST research to ensure that they did not have significant sequence homology to other genes in the database. The design and cloning of shRNAs oligos into PLL3.7 vector were done as described previously (22). The cyt b shRNA oligo consisted of 5′-TGCCGCAGACCTCCTCATTC-TTCAAGAGA-GAATGAGGAGGTCTGCGGC-TTTTTTC (forward). FoxP3 shRNA served as a control. The FoxP3 shRNA oligo was 5′-GGCTTCATCTGTGGCATCA-TTCAAGAGA-TGATGCCACAGATGAAGCCTTTTTTC (forward).

Establishment of HLA-DR1-restricted tumor-specific 1362-D6 CD4+ T cells. Adoptive transfer of TILs together with a high dose of IL-2 has induced objective tumor regressions in melanoma patients (23). TIL1362 cells were generated from a melanoma patient (no. 1,362) who responded to a TIL plus high-dose IL-2 treatment in the Surgery Branch of National Cancer Institute (NCI; Bethesda, MD) and were found to recognize autologous tumor (Fig. 1A). To characterize these T cells further, we used limiting dilution methods to generate T-cell sublines that reacted strongly against autologous tumor cells (data not shown). One of the T-cell sublines, designated 1362-D6, was CD4+ based on FACS analysis (Fig. 1B) and was chosen for antibody-blocking experiments to determine its restriction elements. Recognition of autologous 1362mel cells by 1362-D6 T cells was specifically inhibited by monoclonal antibodies against MHC class II and HLA-DR molecules, but not by antibodies against HLA-DQ, HLA-DP, or MHC class I molecules (Fig. 1C), suggesting that these T cells recognize a tumor antigen presented by HLA-DR molecules. Because 1362mel cells are positive for both HLA-DR1 and HLA-DR11 molecules, we sought to determine the specific HLA-DR restriction element for T-cell recognition. Tumor lysate was pulsed onto EBV-transformed B cells expressing only one of the HLA-DR molecules and used as target cells for T-cell recognition. HLA-DR1-positive 1363 EBV-transformed B cells were able to present antigen specifically from autologous tumor lysate to 1362-D6 CD4+ T cells, whereas autologous tumor lysate-pulsed HLA-mismatched EBV-transformed B cells or allogenic tumor lysate-pulsed HLA-matched 1363 EBV-transformed B cells failed to activate T cells (Fig. 1D). These findings suggest that HLA-DR1 is the restriction element for 1362-D6 CD4+ T-cell recognition and that the tumor antigen is unique to 1362mel cells. To confirm whether the tumor antigen is specifically presented by 1362mel cells, we tested a panel of HLA-DR1-expressing melanoma and EBV-transformed B-cell lines for T-cell recognition and found that 1362-D6 CD4+ T cells specifically recognized autologous 1362mel cells but not HLA-matched or mismatched tumor cells or autologous B cells (Fig. 1E), suggesting that 1362-D6 CD4+ T cells recognize a tumor antigen presented only by 1362mel cells.

Figure 1.

Recognition of the autologous 1362 melanoma cell line by 1362-D6 CD4+ T cells. A, 1362TIL recognized autologous 1362mel cells. B, FACS analysis of 1362-D6 T cells for CD4 expression. C, HLA restriction of T-cell recognition. 1362-D6 T cells were cocultured with autologous 1362mel cells in the presence and absence of various anti-MHC antibodies. D, 1362-D6 CD4+ T cells recognized autologous tumor antigen presented by HLA-DR1 molecules. E, 1362-D6 CD4+ T cells recognized unique tumor antigen.

Figure 1.

Recognition of the autologous 1362 melanoma cell line by 1362-D6 CD4+ T cells. A, 1362TIL recognized autologous 1362mel cells. B, FACS analysis of 1362-D6 T cells for CD4 expression. C, HLA restriction of T-cell recognition. 1362-D6 T cells were cocultured with autologous 1362mel cells in the presence and absence of various anti-MHC antibodies. D, 1362-D6 CD4+ T cells recognized autologous tumor antigen presented by HLA-DR1 molecules. E, 1362-D6 CD4+ T cells recognized unique tumor antigen.

Close modal

Identification of cyt b as the gene encoding a tumor antigen recognized by 1362-D6 CD4+ T cells. To identify the gene that encodes the tumor antigen recognized by 1362-D6 CD4+ T cells, we used a genetic targeting expression system previously with MHC class II–restricted tumor antigens (7). Briefly, a cDNA library was generated from 1362mel cells and fused with a target Ii leading sequence. cDNA subpools with ∼100 cDNA clones per pool were prepared and introduced into the 293IMDR1 APCs expressing DMA, DMB, Ii, DRA, and DRB1*0101 molecules (7). After screening a total of 1.1 × 105 cDNA clones, we identified a positive pool, no. 69, which conferred recognition by 1362-D6 CD4+ T cells when transfected into 293IMDR1 cells (Fig. 2A). Individual cDNA clones were then isolated from the positive pool and tested for recognition by 1362-D6 CD4+ T cells. cDNA clone 2 specifically stimulated 1362-D6 CD4+ T-cell responses, whereas control cDNAs failed to activate T cells (Fig. 2B). DNA sequencing analysis revealed that cDNA clone 2 contained a DNA fragment encoding the COOH-terminal end of mitochondrial cyt b protein, which is part of the respiratory enzyme complex III involved in electron transport (Fig. 2C).

Figure 2.

Screening of an Ii-cDNA library constructed from 1362mel RNA using 1362-D6 CD4+ T cells. A, identification of a positive cDNA pool after screening Ii-cDNA library generated from 1362mel RNA. B, isolation of a positive cDNA clone encoding a tumor antigen recognized by 1362-D6 CD4+ T cells. 1362-D6 CD4+ T cells recognized 293IMDR1 cells transfected with the cDNA clone 2 but not with control cDNA or green fluorescent protein. C, schematic presentation of the nucleotide and amino acid sequences of the Ii-human mitochondrial cyt b fusion cDNAs. Protein residues are inferred from the third ORF of the cyt b gene using nuclear genetic codes. D, four 1362mel-specific CD4+ T-cell sublines recognized 293IMDR1 cells transfected with cDNA clone 2.

Figure 2.

Screening of an Ii-cDNA library constructed from 1362mel RNA using 1362-D6 CD4+ T cells. A, identification of a positive cDNA pool after screening Ii-cDNA library generated from 1362mel RNA. B, isolation of a positive cDNA clone encoding a tumor antigen recognized by 1362-D6 CD4+ T cells. 1362-D6 CD4+ T cells recognized 293IMDR1 cells transfected with the cDNA clone 2 but not with control cDNA or green fluorescent protein. C, schematic presentation of the nucleotide and amino acid sequences of the Ii-human mitochondrial cyt b fusion cDNAs. Protein residues are inferred from the third ORF of the cyt b gene using nuclear genetic codes. D, four 1362mel-specific CD4+ T-cell sublines recognized 293IMDR1 cells transfected with cDNA clone 2.

Close modal

Although cyt b is encoded by the mitochondrial genome and translated according to the mitochondrial code, Ii and the fused cDNA are translated in the cytoplasm. By the nuclear genetic code, we found that the Ii leader sequence was fused to a 76-amino acid polypeptide translated from an alternative ORF of the cyt b gene. To determine the frequency of cyt b–specific T cells, we screened 40 T-cell sublines derived from TIL1362 against 293IMDR1 cells expressing an Ii-cyt b cDNA fragment and found that 4 of 40 T-cell sublines were capable of recognizing the cyt b antigen (Fig. 2D).

Identification of the cyt b epitope recognized by 1362-D6 CD4+ T cells. To identify the epitope from cyt b recognized by 1362-D6 CD4+ T cells, we initially made a series of peptides consisting of 15 to 19 amino acid with a HLA-DR1-binding motif. However, none of them were capable of activating T cells (data not shown). To refine the cyt b DNA region that encodes the T-cell epitope, we made a series of 5′-end and 3′-end constructs of the fusion portion of the DNA fragment and tested for their ability to activate 1362-D6 T cells. 3′-End truncation of the DNA construct encoding the first 43 amino acids abolished T-cell recognition, whereas 5′-end truncation of the construct encoding residues 30 to 60 dramatically diminished T-cell recognition (Fig. 3A). By contrast, constructs encoding the first 60 amino acid or residues 13 to 60 continued to be recognized by T cells. These results suggest that the potential T-cell epitope lies within the 48-amino acid polypeptide extending from amino acid 13 to 60. As shown in Fig. 3B, a peptide of this length strongly activated T cells, whereas the shorter peptides could not. Peptide titration experiments showed that the 48-amino acid peptide could be readily detected by T cells at a peptide concentration of ∼300 nmol/L (Fig. 3C). To further determine the minimal peptide for T-cell recognition, we did addition deletion analysis. Figure 3D shows that the shortest peptide capable of stimulating T cells is a 38-amino acid peptide, although further deletions led to loss of T-cell recognition. These results suggest that the 38-amino acid peptide derived from the translation of an alternative ORF of cyt b cDNA was presented to CD4+ 1362-D6 T cells.

Figure 3.

Identification and characterization of peptides capable of stimulating 1362-D6 CD4+ T cells. A, DNA truncation analysis to define the T-cell epitope. A series of 5′-end and 3′-end deletions of the cyt b cDNA fragment was made by PCR. Truncated DNA fragments were then transfected into 293IMDR1 cells and tested for CD4+ T-cell recognition. B, specific recognition of the 48 amino acids peptide by 1362-D6 CD4+ T cells. A series of overlapping synthetic peptides was pulsed onto HLA-DR1-expressing 1558 EBV-transformed B cells for 90 minutes. After three washes, T cells were added and tested for reactivity based on granulocyte macrophage colony-stimulating factor (GM-CSF) release. C, peptide titration for T-cell recognition. 1558 EBV-transformed B cells were incubated with increasing concentrations of 48-amino acid cyt b peptide and cocultured with 1362-D6 CD4+ T cells overnight. IFN-γ release from T cells was determined with an IFN-γ ELISA kit. D, fine deletion analysis of the cyt b peptide. Experiments were done as in (A).

Figure 3.

Identification and characterization of peptides capable of stimulating 1362-D6 CD4+ T cells. A, DNA truncation analysis to define the T-cell epitope. A series of 5′-end and 3′-end deletions of the cyt b cDNA fragment was made by PCR. Truncated DNA fragments were then transfected into 293IMDR1 cells and tested for CD4+ T-cell recognition. B, specific recognition of the 48 amino acids peptide by 1362-D6 CD4+ T cells. A series of overlapping synthetic peptides was pulsed onto HLA-DR1-expressing 1558 EBV-transformed B cells for 90 minutes. After three washes, T cells were added and tested for reactivity based on granulocyte macrophage colony-stimulating factor (GM-CSF) release. C, peptide titration for T-cell recognition. 1558 EBV-transformed B cells were incubated with increasing concentrations of 48-amino acid cyt b peptide and cocultured with 1362-D6 CD4+ T cells overnight. IFN-γ release from T cells was determined with an IFN-γ ELISA kit. D, fine deletion analysis of the cyt b peptide. Experiments were done as in (A).

Close modal

Cyt b is a tumor antigen recognized by CD4+ T cells. To determine if cyt b is a true tumor antigen expressed naturally in 1362mel cells and recognized by 1362-D6 CD4+ T cells, we made a shRNA construct to knock down expression of the cyt b gene. This construct was cotransfected with Ii-cyt b cDNA into 293IMDR1 target cells and tested for its ability to inhibit 1362-D6 CD4+ T-cell recognition of the cyt b antigen. As shown in Fig. 4A, cyt b–specific shRNA molecules inhibited recognition of Ii-cyt b–expressing 293IMDR1 cells by 1362-D6 CD4+ T cells, in contrast to the control vector, which did not affect recognition. When autologous (1362mel) tumor cells were transduced with lentiviral constructs carrying cyt b shRNAs and tested against 1362-D6 CD4+ T cells, their recognition by the T cells was significantly inhibited (55%), whereas neither the control vector alone nor the FoxP3-specific shRNA (Fig. 4B) had any effect on recognition. To ensure that cyt b shRNA does not have off-target effects, we used CD4+ T cells with different antigen specificity. As shown in Fig. 4B, the 1359mel-specific CD4+ T-cell recognition of 1359mel target cells was not inhibited when the target cells were infected with a lentivirus-encoding cyt b shRNA, or FoxP3-specific shRNA was not inhibited. These results suggest that cyt b is a true tumor antigen expressed in 1362mel cells recognized by 1362-D6 CD4+ T cells.

Figure 4.

Validation of cyt b as a natural tumor antigen in 1362mel cells. A, cyt b–specific shRNA expressed from a lentiviral expression vector inhibited cyt b antigen presentation. Lentiviral vector expressing a cyt b–specific shRNA was cotransfected with Ii-cyt b cDNA into 293IMDR1 cells and tested for 1362-D6 CD4+ T-cell recognition. B, inhibition of 1362-D6 CD4+ T-cell recognition of autologous 1362mel cells by cyt b–specific shRNA. Lentivirus particles carrying cyt b–specific shRNA were used to infect 1362mel cells. Three days after infection, tumor cells were harvested and tested for CD4+ T-cell recognition. FoxP3-specific shRNA served as a negative control. 1359 CD4+ T-cell recognition of 1359mel cells was used to show the specificity of cyt b–specific shRNA.

Figure 4.

Validation of cyt b as a natural tumor antigen in 1362mel cells. A, cyt b–specific shRNA expressed from a lentiviral expression vector inhibited cyt b antigen presentation. Lentiviral vector expressing a cyt b–specific shRNA was cotransfected with Ii-cyt b cDNA into 293IMDR1 cells and tested for 1362-D6 CD4+ T-cell recognition. B, inhibition of 1362-D6 CD4+ T-cell recognition of autologous 1362mel cells by cyt b–specific shRNA. Lentivirus particles carrying cyt b–specific shRNA were used to infect 1362mel cells. Three days after infection, tumor cells were harvested and tested for CD4+ T-cell recognition. FoxP3-specific shRNA served as a negative control. 1359 CD4+ T-cell recognition of 1359mel cells was used to show the specificity of cyt b–specific shRNA.

Close modal

Cyt b transcript is derived from the mitochondrion. Because the T-cell peptide was translated in the cytoplasm using nuclear genetic code, we sought to determine the cellular localization of the translated products. Figure 5A shows that protein fractions derived from tumor cell cytoplasm and lysosomes were capable of stimulating 1362-D6 CD4+ T cells when pulsed on HLA-DR1-expressing EBV-transformed B cells, whereas proteins fractions of mitochondria failed to stimulate T cells. These results suggest that cyt b T-cell peptides were translated and localized in the cytoplasm. We next sought to determine how cyt b mRNA was translocated to the cytoplasm of tumor cells. One possibility is that the cyt b gene was transferred and integrated into the nuclear genome. However, our Southern blot analysis using multiple restriction enzyme digestions and tumor cell lines failed to identify a cyt b DNA fragment unique to 1362mel (data not shown), suggesting that cyt b gene in 1362mel might not be integrated into the nuclear genome of 1362mel cells. Alternatively, the cyt b transcript could be transcribed in the mitochondria but leaked out to the cytoplasm for translation. To test this possibility, we treated tumor cells with mitochondrion inhibitor rhodamine 6G for 2 days and then tested for T-cell recognition. Figure 5B shows that rhodamine 6G specifically inhibited 1362-D6 CD4+ T-cell recognition of 1362mel cells, whereas control CD4+ T cells that recognize nuclear tumor antigens were not affected, suggesting that mitochondrion is the source of cyt b transcript that encodes the T-cell peptide. To determine if the cyt b protein was processed using the MHC class II pathway, we tested 1362-D6 CD4+ T-cell recognition of 1362mel cells in the presence of chloroquine inhibitor, a lysosomotropic agent that inhibits MHC class II antigen processing. As shown in Fig. 5C, the addition of chloroquine inhibitor specifically inhibited 1362-D6 CD4+ T-cell recognition of 1362mel cells, whereas little inhibitory effects were observed with control CD8+ T cells that recognize MHC class I–restricted EBNA1 protein (24). Taken together, these results suggest that cyt b mRNA was leaked out of the mitochondrion for translation in the cytoplasm, which is then processed and presented to CD4+ T cells via the MHC class II antigen presentation pathway.

Figure 5.

Identification of mitochondrion as the source of cyt b transcript. A, cyt b peptide was synthesized and localized in the cytoplasm. Protein fractions were isolated from organelles and pulsed onto autologous 1362 EBV-transformed B cells. Protein lysates from 586mel served as negative control. B, specific inhibition of 1362-D6 T-cell recognition of 1362mel by mitochondrion inhibitor. Tumor cells were treated with various concentrations of rhodamine 6G for 2 days. After washing, cells were incubated with T cell overnight for IFN-γ release assays. C, blocking of 1362-D6 T-cell recognition of 1362mel by MHC class II antigen-processing inhibitor chloroquine. CD8+ T-cell recognition of EBNA1 served as negative control.

Figure 5.

Identification of mitochondrion as the source of cyt b transcript. A, cyt b peptide was synthesized and localized in the cytoplasm. Protein fractions were isolated from organelles and pulsed onto autologous 1362 EBV-transformed B cells. Protein lysates from 586mel served as negative control. B, specific inhibition of 1362-D6 T-cell recognition of 1362mel by mitochondrion inhibitor. Tumor cells were treated with various concentrations of rhodamine 6G for 2 days. After washing, cells were incubated with T cell overnight for IFN-γ release assays. C, blocking of 1362-D6 T-cell recognition of 1362mel by MHC class II antigen-processing inhibitor chloroquine. CD8+ T-cell recognition of EBNA1 served as negative control.

Close modal

High percentage of mutations in 1362mel cells. Because 1362-D6 CD4+ T cells recognized 1362mel, but not HLA-DR1-matched tumor cells or EBV-transformed B cells, we reasoned that the cyt b gene in 1362mel tumor cells might contain mutations. To test this possibility, we sequenced >400 independent cyt b cDNA clones from 1362mel and HLA-DR1-matched melanoma cell lines (Table 1). Figure 5 shows extensive DNA mutations resulting in a change of cyt b residues specific for 1362mel cells, although only a few mutations were observed in autologous CD4+ T cells and other control melanoma cell lines. More than 50% of the cDNA clones sequenced contained mutations that were not seen in control cells and resulted in a 37% change in amino acid residues (Table 1). The majority of the mutations was heteroplasmic, whereas only three mutations were homoplasmic. One of the three homoplasmic mutations resulted in an amino acid change from L to I (residue 236) that was associated with ischemic cardiomyopathy (Fig. 6; ref. 25). Heteroplasmic missense mutants (5% of total cDNA clones) were also detected on W113R and W326R residues that were associated with exercise intolerance (26). A majority (81%) of the heteroplasmic mutations were transition substitutions (T→C, 38% and A→G, 35%; Table 2), suggesting that T→C and A→G are common mutations in cyt b. This mutational spectrum is consistent with known hotspots from mitochondrial mutations commonly found in normal and tumor cells (2730).

Table 1.

Summary of cyt b DNA and amino acid residue changes in cell lines

Cell line, n (%)
1362mel1362 T cells1558mel1558 EBV1363mel
cDNA clone sequenced 192 80 96 48 96 
Mutations 117 (60%) 9 (11%) 8 (8%) 4 (8%) 8 (8%) 
Residue changes 71 (37%) 3 (4%) 4 (4%) 1 (3%) 3 (3%) 
Cell line, n (%)
1362mel1362 T cells1558mel1558 EBV1363mel
cDNA clone sequenced 192 80 96 48 96 
Mutations 117 (60%) 9 (11%) 8 (8%) 4 (8%) 8 (8%) 
Residue changes 71 (37%) 3 (4%) 4 (4%) 1 (3%) 3 (3%) 

NOTE: Mutations were confirmed by sequencing from both DNA strands.

Figure 6.

Identification of extensive mutations unique to 1362mel cells. Mitochondrial cyt b protein sequences of 200 cDNA clones derived from 1362mel cells were aligned with those obtained from autologous normal tissue (1362 CD4+ T cells) and HLA-DR1-expressing control melanoma cell lines (Table 1). Amino acid residues were inferred from the DNA sequences with the use of vertebrate mitochondrial codes. Protein sequences from the 1558mel cell line were used as a reference. Protein residue changes within each cell line were inferred from all of the cDNA clones sequenced. Mutant L236I, associated with ischemic cardiomyopathy, is denoted by “*” above the residue. Nonsense mutations, detected in a few cyt b cDNA clones from 1362mel and 1558mel cells, are indicated with an “x.”

Figure 6.

Identification of extensive mutations unique to 1362mel cells. Mitochondrial cyt b protein sequences of 200 cDNA clones derived from 1362mel cells were aligned with those obtained from autologous normal tissue (1362 CD4+ T cells) and HLA-DR1-expressing control melanoma cell lines (Table 1). Amino acid residues were inferred from the DNA sequences with the use of vertebrate mitochondrial codes. Protein sequences from the 1558mel cell line were used as a reference. Protein residue changes within each cell line were inferred from all of the cDNA clones sequenced. Mutant L236I, associated with ischemic cardiomyopathy, is denoted by “*” above the residue. Nonsense mutations, detected in a few cyt b cDNA clones from 1362mel and 1558mel cells, are indicated with an “x.”

Close modal
Table 2.

Summary of cyt b mutations in 1362mel

DNA substitutionNo. mutationsPercentage of total
T→C 45 38 
A→G 41 35 
C→T 
G→A 
A→C 
T→G 
C→A 
A→T 
DNA substitutionNo. mutationsPercentage of total
T→C 45 38 
A→G 41 35 
C→T 
G→A 
A→C 
T→G 
C→A 
A→T 

NOTE: Number of mutation refers to total mutations obtained from all cDNA clones sequenced.

This report describes for the first time the identification of mitochondrial-derived cyt b as a tumor antigen recognized by 1362-D6 CD4+ T cells. We showed its antigenic authenticity in 1362mel cells using gene knockdown experiments with cyt b–specific shRNA (Fig. 4B). We further showed by 5′-end and 3′-end deletion analysis that the epitope recognized by CD4+ T cells is derived from a 38-amino acid cyt b peptide (Fig. 3). However, the 38-amino acid peptide is not the final natural peptide presented on the tumor surfaces because it could not be presented to CD4+ T cells when pulsed onto HLA-DR1-expressing tumor cells or 293IMDR1 cells with poor endocytosis capability (data not shown). Further studies are needed to elucidate the mechanism by which natural cyt b peptides are generated for T-cell recognition.

Our data favor the hypothesis that the cyt b mRNA is leaked out of the mitochondrion and then translated in the cytoplasm. We first showed that the cyt b peptide was derived from the cytoplasm and not from mitochondrion because protein lysate isolated from mitochondria was incapable of stimulating T cells when pulsed on APCs (Fig. 5A). This agrees with the Ii-cyt b cDNA that encodes the cyt b peptide using nuclear genetic code (Fig. 2C). However, we also showed that cyt b–specific T-cell recognition of autologous tumor cells was effectively inhibited by rhodamine 6G, which depletes mitochondrion (Fig. 5B), suggesting that mitochondrion is the source of cyt b transcript. These results are consistent with our Southern blot analysis using four restriction enzymes and four melanoma cell lines that failed to detect a cyt b DNA fragment unique to 1362mel cells (data not shown), suggesting that the cyt b gene is not transferred and integrated to the nucleus.

Mitochondrial cyt b is part of the respiratory enzyme complex III, which catalyzes electron transfer from succinate and NAD+-linked dehydrogenases to cytochrome c. Mutations in mitochondrial cyt b have been associated with several human diseases, including progressive exercise intolerance and complex III deficiency (31), myoglobinuria (32), Leber's hereditary optic neuropathy (33), ischemic cardiomyopathy (25), parkinsonism/mitochondrial encephalopathy, lactate acidosis, and stroke-like episodes (34), and severe hypertrophic cardiomyopathy (35). All of these reported cases of cyt b mutations were heteroplasmic (∼70-80% mutant genes in affected organs only), suggesting that they were somatic in origin. By contrast, the C→A mutation found at position 15452 in our patient's cells, converting residue 236 of cyt b from leucine to isoleucine (25), was detected in 100% of the cyt b cDNA clones derived from 1362mel and 1362 T cells, suggesting a germ-line origin. Because patients carrying this mutation had a reduction in respiratory complex III activity (>50%), the mutation may play a role in the pathogenesis or progression of ischemic cardiomyopathy (25). The L236 residue, although highly conserved in many organisms, is located within transmembrane E, a region of the protein not associated with haem or quinone binding. Thus, the L236I mutation is unlikely to have an adverse effect on the bc1 complex (36), which may explain the prolonged survival of our patient whose mutation of this residue was homoplasmic.

Mitochondrial genome is useful for studies of human evolution, migration, and population histories because of its high copy number, maternal inheritance, lack of recombination, and a generally higher mutation rate than found in nuclear DNA (37). However, the accumulation of reactive oxygen species in some cells over time could induce a high level of mutations in both nuclear and mitochondrial DNAs (mtDNA), causing significant damage to mitochondrial and cellular proteins, lipids, and nucleic acids, which may predispose the host to cancer (38). Indeed, in our melanoma patient, ∼37% of the cyt b protein in the mitochondria were mutated (Table 1), suggesting that these changes and perhaps others still to be identified in patient's mtDNAs may contribute either directly or indirectly to tumorigenesis. Given that extensive mutations occurred in cyt b in tumor cells, we investigated whether mutations in cyt b contributed to recognition of cyt b antigen by T cells. We screened at least 20,000 cyt b cDNA derived from 1362mel using our 293IMDR1 cells as APCs. However, we failed to identify specific mutations responsible for T-cell recognition (data not shown). One possible explanation is that tumor cells may have the ability to more efficiently process and present T-cell epitopes than 293IMDR1 cells. Nonetheless, although our results indicated that the T-cell epitope encoded by a cyt b cDNA fragment did not contain mutations, mutations in other regions of cyt b may collectively contribute to aberrant expression of cyt b alternative gene products, which may directly or indirectly affect the immunogenicity of tumor.

Although cutaneous malignant melanoma has been convincingly linked with exposure to UV radiation (39, 40), few specific genetic mutations have been implicated in its induction and biological behavior (16). Our study provides the first example of the association of melanoma with cyt b mutations. Although the tumorigenic contribution of these mutations remains to be determined, it should be noted that an L236I homoplasmic mutation in a patient with decreased complex III activity is likely to increase the mitochondrial oxidative stress and activate nuclear DNA proto-oncogenes that could result in neoplastic transformation (41, 42). Most importantly, perhaps, our findings provide the first example of a mitochondrial immune target for CD4+ T cells and thus have potential implications for the immunosurveillance of mitochondrial aberrations associated with cancer.

Note: K.S. Voo and G. Zeng contributed equally to this work.

Grant support: NIH grants R01 CA101795, R01 CA90327, and P50 CA093459 (R-F. Wang).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Steven A. Rosenberg (Surgery Branch, NCI, NIH) for providing TILs and melanoma cell lines and Dr. Steve Zullo (Laboratory of Biochemical Genetics, National Institute of Mental Health, NIH, Bethesda, MD) for helpful discussions on mitochondria fractionation.

1
Wang RF. The role of MHC class II-restricted tumor antigens and CD4+ T cells in antitumor immunity.
Trends Immunol
2001
;
22
:
269
–76.
2
Janssen EM, Lemmens EE, Wolfe T, Christen U, von Herrath MG, Schoenberger SP. CD4+ T cells are required for secondary expansion and memory in CD8+ T lymphocytes.
Nature
2003
;
421
:
852
–6.
3
Shedlock DJ, Shen H. Requirement for CD4 T cell help in generating functional CD8 T cell memory.
Science
2003
;
300
:
337
–9.
4
Sun JC, Bevan MJ. Defective CD8 T cell memory following acute infection without CD4 T cell help.
Science
2003
;
300
:
339
–42.
5
Fu T, Voo KS, Wang RF. Critical role of EBNA1-specific CD4+ T cells in the control of mouse Burkitt lymphoma in vivo.
J Clin Invest
2004
;
114
:
542
–50.
6
Wang HY, Fu T, Wang G, et al. Induction of CD4+ T cell-dependent antitumor immunity by TAT-mediated tumor antigen delivery into dendritic cells.
J Clin Invest
2002
;
109
:
1463
–70.
7
Wang RF, Wang X, Atwood AC, Topalian SL, Rosenberg SA. Cloning genes encoding MHC class II-restricted antigens: mutated CDC27 as a tumor antigen.
Science
1999
;
284
:
1351
–4.
8
Schultz-Thater E, Noppen C, Gudat F, et al. NY-ESO-1 tumour associated antigen is a cytoplasmic protein detectable by specific monoclonal antibodies in cell lines and clinical specimens.
Br J Cancer
2000
;
83
:
204
–8.
9
Raposo G, Tenza D, Murphy DM, Berson JF, Marks MS. Distinct protein sorting and localization to premelanosomes, melanosomes, and lysosomes in pigmented melanocytic cells.
J Cell Biol
2001
;
152
:
809
–24.
10
Wang HY, Zhou J, Zhu K, Riker AI, Marincola FM, Wang RF. Identification of a mutated fibronectin as a tumor antigen recognized by CD4+ T cells: its role in extracellular matrix formation and tumor metastasis.
J Exp Med
2002
;
195
:
1397
–406.
11
Shimoda S, Van de Water J, Ansari A, et al. Identification and precursor frequency analysis of a common T cell epitope motif in mitochondrial autoantigens in primary biliary cirrhosis.
J Clin Invest
1998
;
102
:
1831
–40.
12
Matsumura S, Kita H, He XS, et al. Comprehensive mapping of HLA-A0201-restricted CD8 T-cell epitopes on PDC-E2 in primary biliary cirrhosis.
Hepatology
2002
;
36
:
1125
–34.
13
Fan R, Tykodi SS, Braciale TJ. Recognition of a sequestered self peptide by influenza virus-specific CD8+ cytolytic T lymphocytes.
J Immunol
2000
;
164
:
1669
–80.
14
Sasada T, Ghendler Y, Neveu JM, Lane WS, Reinherz EL. A naturally processed mitochondrial self-peptide in complex with thymic MHC molecules functions as a selecting ligand for a viral-specific T cell receptor.
J Exp Med
2001
;
194
:
883
–92.
15
Berg RE, Princiotta MF, Irion S, Moticka JA, Dahl KR, Staerz UD. Positive selection of an H2-3 restricted T cell receptor.
Immunity
1999
;
11
:
33
–3.
16
Poetsch M, Dittberner T, Petersmann A, Woenckhaus C. Mitochondrial DNA instability in malignant melanoma of the skin is mostly restricted to nodular and metastatic stages.
Melanoma Res
2004
;
14
:
501
–8.
17
Voo KS, Fu T, Heslop HE, Brenner MK, Rooney CM, Wang RF. Identification of HLA-DP3-restricted peptides from EBNA1 recognized by CD4+ T cells.
Cancer Res
2002
;
62
:
7195
–9.
18
Wang RF, Johnston SL, Zeng G, Topalian SL, Schwartzentruber DJ, Rosenberg SA. A breast and melanoma-shared tumor antigen: T cell responses to antigenic peptides translated from different open reading frames.
J Immunol
1998
;
161
:
3598
–606.
19
Wang RF, Appella E, Kawakami Y, Kang X, Rosenberg SA. Identification of TRP-2 as a human tumor antigen recognized by cytotoxic T lymphocytes.
J Exp Med
1996
;
184
:
2207
–16.
20
Mu J, Duan J, Makova KD, et al. Chromosome-wide SNPs reveal an ancient origin for Plasmodium falciparum.
Nature (Lond)
2002
;
418
:
323
–6.
21
Aronson NN, Jr., Touster O. Isolation of rat liver plasma membrane fragments in isotonic sucrose.
Methods Enzymol
1974
;
31
:
90
–102.
22
Rubinson DA, Dillon CP, Kwiatkowski A, et al. A lentivirus-based system to functionally silence genes in primary mammalian cells, stem cells, and transgenic mice by RNA interference.
Nat Genet
2003
;
33
:
401
–6.
23
Rosenberg SA, Yannelli JR, Yang JC, et al. Treatment of patients with metastatic melanoma with autologous tumor-infiltrating lymphocytes and interleukin 2.
J Natl Cancer Inst
1994
;
86
:
1159
–66.
24
Voo KS, Fu T, Wang HY, et al. Evidence for the presentation of major histocompatibility complex class I-restricted Epstein-Barr virus nuclear antigen 1 peptides to CD8+ T lymphocytes.
J Exp Med
2004
;
199
:
459
–70.
25
Marin-Garcia J, Hu Y, Ananthakrishnan R, Pierpont ME, Pierpont GL, Goldenthal MJ. A point mutation in the cytb gene of cardiac mtDNA associated with complex III deficiency in ischemic cardiomyopathy.
Biochem Mol Biol Int
1996
;
40
:
487
–95.
26
Andreu AL, Hanna MG, Reichmann H, et al. Exercise intolerance due to mutations in the cytochrome b gene of mitochondrial DNA.
N Engl J Med
1999
;
341
:
1037
–44.
27
Nekhaeva E, Bodyak ND, Kraytsberg Y, et al. Clonally expanded mtDNA point mutations are abundant in individual cells of human tissues.
Proc Natl Acad Sci U S A
2002
;
99
:
5521
–6.
28
Coller HA, Khrapko K, Bodyak ND, Nekhaeva E, Herrero-Jimenez P, Thilly WG. High frequency of homoplasmic mitochondrial DNA mutations in human tumors can be explained without selection.
Nat Genet
2001
;
28
:
147
–50.
29
Polyak K, Li Y, Zhu H, et al. Somatic mutations of the mitochondrial genome in human colorectal tumours.
Nat Genet
1998
;
20
:
291
–3.
30
Khrapko K, Coller HA, Andre PC, Li XC, Hanekamp JS, Thilly WG. Mitochondrial mutational spectra in human cells and tissues.
Proc Natl Acad Sci U S A
1997
;
94
:
13798
–803.
31
Borisov VB. Defects in mitochondrial respiratory complexes III and IV and human pathologies.
Mol Aspects Med
2002
;
23
:
385
–412.
32
Andreu AL, Bruno C, Dunne TC, et al. A nonsense mutation (G15059A) in the cytochrome b gene in a patient with exercise intolerance and myoglobinuria.
Ann Neurol
1999
;
45
:
127
–30.
33
Johns DR, Neufeld MJ. Cytochrome b mutations in Leber hereditary optic neuropathy.
Biochem Biophys Res Commun
1991
;
181
:
1358
–64.
34
De Coo IF, Renier WO, Ruitenbeek W, et al. A 4-base pair deletion in the mitochondrial cytochrome b gene associated with parkinsonism/MELAS overlap syndrome.
Ann Neurol
1999
;
45
:
130
–3.
35
Valnot I, Kassis J, Chretien D, et al. A mitochondrial cytochrome b mutation but no mutations of nuclearly encoded subunits in ubiquinol cytochrome c reductase (complex III) deficiency.
Hum Genet
1999
;
104
:
460
–6.
36
Fisher N, Meunier B. Effects of mutations in mitochondrial cytochrome b in yeast and man. Deficiency, compensation, and disease.
Eur J Biochem
2001
;
268
:
1155
–62.
37
Pakendorf B, Stoneking M. Mitochondrial DNA and human evolution.
Annu Rev Genomics Hum Genet
2005
;
6
:
165
–83.
38
Wallace DC. Mitochondrial diseases in man and mouse.
Science
1999
;
283
:
1482
–8.
39
Holman CD, Armstrong BK, Heenan PJ. A theory of the etiology and pathogenesis of human cutaneous malignant melanoma.
J Natl Cancer Inst
1983
;
71
:
651
–6.
40
Whiteman DC, Whiteman CA, Green AC. Childhood sun exposure as a risk factor for melanoma: a systematic review of epidemiologic studies.
Cancer Causes Control
2001
;
12
:
69
–821.
41
Wallace DC. Mouse models for mitochondrial disease.
Am J Med Genet
2001
;
106
:
71
–93.
42
Eng C, Kiuru M, Fernandez MJ, Aaltonen LA. A role for mitochondrial enzymes in inherited neoplasia and beyond.
Nat Rev Cancer
2003
;
3
:
193
–202.